Joint Injury, Osteoarthritis, and Cardiovascular Disease Risk Factors in Former National Football League Athletes: An NFL-LONG Study

Former NFL players completed a general health questionnaire either online (Qualtrics; SAP America Inc) or via hard copy. This cross-sectional study was approved by the Institutional Review Board at The University of North Caroline at Chapel Hill, and all recruits provided written informed consent before participation.

More than 15 000 former NFL players were contacted via mail, and a subset of these same individuals (n = 11 645) were also contacted via electronic mail (Figure). Former players were eligible to participate if they had played ≥1 full seasons in the NFL. Recruitment took place from January 1, 2019, to February 14, 2020, and data analysis took place from August 11, 2020, to May 17, 2021.

The comprehensive questionnaire we used has been described elsewhere as part of an ongoing study: Neurologic Function Across the Lifespan: A Prospective, Longitudinal, and Translational Study for Former National Football League Players (NFL-LONG).^23,24  This survey acquires information about former players’ personal demographics, football playing history, medical history, concussion history, musculoskeletal injury and surgery history, self-reported functioning, health-related quality of life, current substance use, and health-related behaviors.

Outcomes of interest were self-reported lifetime medical diagnoses of CVD (coronary heart disease, myocardial infarction, and peripheral vascular disease) and, separately, CRFs (hypertension, hypercholesterolemia, diabetes mellitus type II, sleep apnea, chronic obstructive pulmonary disease). For each diagnosis, participants were asked, “Have you ever been told by a physician or health professional that you had/have any of the following conditions?”

Lifetime medical diagnosis of OA was ascertained using the same question as above for the CVD and CRF conditions. Before completing questions about musculoskeletal injury and surgical histories, participants were provided the following definition: “Musculoskeletal injuries are any injury to the muscles, tendons, ligaments, cartilage, and bones. They can include, but are not limited to sprains, strains, dislocations and fractures.” Next, participants were asked if they had ever sustained a musculoskeletal injury when they played football. Individuals with a history of a musculoskeletal injury were also asked to indicate the number of lifetime injuries and surgeries experienced for various body regions. For the present study, only knee, hip, and ankle injuries, surgeries, and OA were examined.

Age, race and ethnicity, obesity status, lifetime medical diagnosis of liver or kidney disease, and smoking history were included as covariates. Age was calculated as the number of whole years between the participant’s date of birth and the date of survey completion. Racial or ethnic identities were self-reported. To obtain obesity status, each participant’s BMI (kg/m²) was calculated from the self-reported height and weight, and then the individual was classified as either not obese (BMI < 30) or obese (BMI ≥ 30). Liver and kidney disease diagnoses were acquired using the same question shown earlier for CVD, CRF, and OA diagnoses. Smoking history consisted of a yes or no response to the question, “Have you smoked 100 or more cigarettes in your lifetime? (Note: There are 20 cigarettes in one pack.)”

We used a single contrast predictor variable with 3 categories for injury, surgery, and OA history: (1) those with no injury or surgical history and no OA diagnosis (neither condition was reported; referent category), (2) those with a history of ≥1 lower extremity joint injuries or surgeries but no OA diagnosis (reported injury or surgery only), and (3) those with a history of ≥1 lower extremity joint injuries or surgeries and a diagnosis of OA (both conditions were reported).

To calculate prevalence ratios (PRs) for CVD, we computed a bivariate log-binomial regression model for the injury, surgery, and OA predictor variable. Next, a multivariable log-binomial regression model was fit to calculate the adjusted PRs (PRs_adj) for the predictor contrast variables while covarying for age (continuous), racial or ethnic identity (operationalized dichotomously based on the sample distribution: those who identified as White or non-Hispanic and those who did not), obesity status (obese versus not obese), and liver or kidney disease (dichotomous; lifetime diagnosis of liver disease, kidney disease, or both versus neither diagnosis). Poisson regression models were conducted to determine PRs for the CRF outcome variable, as the log-binomial regression models failed to provide stable solutions.²⁵  Empirical Wald tests based on the robust sandwich covariance estimator were used to evaluate statistical significance, and effect estimates with 95% CIs excluding 1.0 were considered statistically significant. In post hoc exploratory analyses, we separated the study participants into 2 groups based on age (those <50 years of age and those ≥50 years of age) and fit the same models to each group separately.

Our secondary aim was to describe the knee-specific injury and surgery history in relation to CVD and CRF prevalence. These analyses mirrored those for the primary aim with the full sample but did not include participants who did not report a history of knee-related injury or surgery. All analyses were performed in SPSS (version 25; IBM Corp).

As of February 2020, a total of 1784 former NFL football players had completed the initial questionnaire. Among those respondents, 98 (5.5%) did not provide a response to ≥1 variables of interest in the present study: injury or surgery information (n = 18), racial or ethnic identity (n = 11), height or weight (from which we calculated BMI; n = 17), and smoking history (n = 56). Only 6% (n = 18) of participants reported an OA diagnosis but no history of lower extremity joint injury or surgery; this group was excluded from our analyses. Thus, 1686 former players were included in the current analyses (Table 1), of whom more than half had a history of lower extremity surgery (Table 2).

Regarding CVD diagnoses, 145 participants in this sample (8.6%) reported a diagnosis of coronary heart disease or heart attack but not peripheral vascular disease, 17 (1.0%) indicated a diagnosis of peripheral vascular disease but not coronary heart disease or heart attack, and 9 (0.5%) cited both diagnoses. A total of 899 participants (53.3%) reported ≥1 CRF. More specifically, 458 (27.2%) described only 1 diagnosis, 291 (17.3%) indicated 2 diagnoses, 131 (7.8%) cited 3 of the diagnoses, and 19 (1.1%) reported having all 4 CRF diagnoses (hypertension, hypercholesterolemia, type II diabetes, and sleep apnea).

Neither a history of lower extremity injury or surgery without OA nor a history of both lower extremity injury or surgery and diagnosed OA was significantly associated with CVD after controlling for age, racial or ethnic identity, obesity status, and diagnoses of liver disease, kidney disease, or both when compared with those reporting no prior injuries, surgeries, or OA diagnosis (Table 3). This lack of a relationship with CVD prevalence remained even when participants were separated into older and younger age groups (ie, those ≥50 years of age and those <50 years old, respectively).

Conversely, the prevalence of reporting a diagnosis of ≥1 CRFs was higher among former players with a history of lower extremity injury or surgery with or without diagnosed OA versus athletes with neither condition (Table 4; PRs adjusted for the prevalence of each individual CRF diagnosis are presented in Supplemental Table 1, available online at http://dx.doi.org/10.4085/1062-6050-0437.22.S1). When relationships with the prevalence of any CRF were explored by age group, participants ≥50 years who had both a history of injury or surgery and a diagnosis of OA had a higher prevalence of ≥1 CRFs compared with those who had neither diagnosis, but this was not true for those with a history of injury or surgery without an OA diagnosis. Younger participants (<50 years of age) exhibited significant PRs in both injury and OA contrast conditions compared with those who had no injury, surgery, or OA diagnosis. Further, the magnitude of the PR_adj for CRFs in younger participants with an injury or surgery history and a diagnosis of OA was twice as large as the PR_adj for CRFs in older participants.

Finally, we repeated our primary analyses (excluding exploratory models by age) only in participants who had a history of injury or surgery to the knee with or without OA anywhere in the lower extremity. Of note, former players with OA were approximately 20 years older than former players without OA on average (Supplemental Table 2). Former players with a history of knee injury or surgery, whether or not they reported OA, had a higher prevalence of reporting a CVD diagnosis after we controlled for age, racial or ethnic identity, obesity status, and diagnosis of liver disease, kidney disease, or both (Supplemental Table 3). However, CRFs were more prevalent among former players with a history of knee injury or surgery and those with a history of knee injury or surgery and diagnosed OA versus those with no history of either condition (Supplemental Table 4).

Risk factors for CVD are more prevalent among former NFL players with a history of lower extremity injury or surgery. The PRs reported in Table 4 indicate that the association between injury or disease status and CRFs appears to be consistently higher in individuals with a history of injury or surgery and OA as compared with individuals who had a history of injury or surgery without OA; however, it should be noted that the CIs of these estimates overlap in all cases. More specifically, lower extremity joint injury or surgery alone was associated with a 30% greater prevalence of ≥1 CRFs, and individuals with lower extremity joint injury or surgery and OA reported a 53% greater prevalence of CRFs versus former players without a history of lower extremity joint injury, surgery, or OA. These findings are consistent with the literature describing an association of knee and hip OA with both hypercholesteremia and obesity status in other populations.²⁶ 

Elevated CRFs (eg, elevated systolic blood pressure) have been described among former athletes, and an elevated risk of CVD among former athletes was associated with an elevated BMI.²⁷  Our study furthers this work by demonstrating a greater risk of CRFs among former athletes with a history of lower extremity joint injury or surgery in the absence of OA. Although our research did not allow us to establish causation or the mechanisms underlying this association, several groups have indicated that individuals with a history of lower extremity joint injury or surgery had more joint pain,²⁸  poorer cardiorespiratory fitness,^29,30  greater adiposity^29,31  and participated in less physical activity^13,14  than their peers who had not experienced an injury. Athletic trainers are uniquely positioned within sports organizations to implement evidence-based lower extremity injury-prevention strategies and provide comprehensive rehabilitation for lower extremity joint injuries; however, our results signify that an enhanced focus on patient counseling about risk factors for chronic disease development (eg, OA and CVD) after lower extremity injury may be warranted to improve patient well-being across the lifespan.

A history of lower extremity injury or surgery either in isolation or in combination with diagnosed OA was not significantly associated with the prevalence of CVD among this sample of former NFL athletes. We operationally defined CVD as the presence of coronary heart disease, myocardial infarction, or peripheral vascular disease, which is an important distinction from previous studies of former NFL players whose authors limited their focus to myocardial infarction.¹¹  In a meta-analysis of the association between OA and CVD, defined as coronary and ischemic heart disease, myocardial infarction, congestive heart failure, and stroke, individuals who had knee or hip OA experienced a 30% or 23% greater relative risk of CVD, respectively, as compared with healthy individuals.¹⁸  Our PR_adj showed that individuals with a history of lower extremity joint injury or surgery and OA had a 41% greater prevalence of CVD (95% CI = 0.89, 2.25) and individuals with a history of lower extremity joint injury or surgery but no diagnosed OA had a 34% greater prevalence versus former players with no history of injury, yet these findings were not statistically significant and should therefore be interpreted with caution. The authors of a recent meta-analysis of CVD and CRFs in former athletes found that the prevalence of CVD was comparable with that of the nonathlete population, indicating that postcareer changes in health behaviors may result in a cardiovascular risk profile that is not unlike that of nonathletes.²⁷  Future studies would benefit from enhanced detail regarding health behaviors (eg, participation in physical activity) that may be negatively affected by joint injury,^13,14  OA,¹⁶  and the end of the playing career to allow us to better understand the role of joint health during the lifespan of former NFL players.

Older individuals are at higher risk for CRFs and CVD.³²  Neither a history of lower extremity injury or surgery without OA nor a history of lower extremity injury or surgery and diagnosed OA were significantly associated with CVD in the older or younger age groups. However, younger individuals (<50 years old) with a history of lower extremity injury or surgery without OA experienced a 53% greater prevalence of CRFs and individuals with a history of both lower extremity injury or surgery and OA experienced a 146% greater prevalence of CRFs when compared with younger individuals who had no history of injury, surgery, or OA (Table 4). Alternatively, older individuals (≥50 years old) who had both a history of injury or surgery and a diagnosis of OA displayed a 20% greater prevalence of CRFs as compared with individuals who had no history of injury, surgery, or OA, whereas injury or surgery in the absence of OA was not significantly associated with a greater prevalence of CRFs. This finding seems counterintuitive, as the prevalence of CRFs is known to increase with age.³³  Nonetheless, older individuals in our cohort were more likely to develop CRFs and CVD independent of injury, surgery, or OA status. Thus, the greater prevalence of CRFs observed before the typical period of highest CVD risk (<50 years old) may reflect an important pathway from lower extremity injury to premature morbidity that requires additional longitudinal examination.

Our knee-specific analysis mirrored our omnibus analysis, which simultaneously considered hip-, knee-, and ankle-joint injuries. This is likely because 91.1% of individuals who reported a diagnosis of OA in the sample stated that their diagnosis involved the knee joint, and 71.6% of individuals who recounted a history of lower extremity joint injury or surgery described ≥1 knee injuries or surgeries. These outcomes are similar to those of previous investigators^6,10  who determined that knee-joint surgery was the most common lower extremity surgery and knee-related posttraumatic OA the most prevalent OA diagnosis among NFL athletes as compared with nonathletes who had similar demographics.

This was a cross-sectional subanalysis of a larger ongoing cohort study. Although the sample of former NFL players in this study was large (N = 1686), we were not able to establish if meaningful differences in demographics or health history existed between individuals who did and those who did not respond to the survey. This may have introduced sampling and response bias, both of which might have affected the findings. Given the cross-sectional design, we were not able to identify the incidence of CRFs or CVD in our sample, which limited us to reporting the association between injury history and CRF or CVD prevalence. This also restricted our ability to characterize the timing of the injury diagnosis or OA diagnosis relative to the playing career, as well as any change in behavior (eg, change in physical activity level or cessation of smoking) that might have influenced CRFs or the risk of CVD. Health histories were self-reported, and we were unable to confirm responses via chart review or other sources. This is particularly concerning because many of the medical conditions included in the health history can be described using several terms (eg, medical versus layperson), which may have caused confusion among some respondents. Lastly, it is important to consider the generalizability of this study. American football is a unique sport, and the individuals who play football at the NFL level may differ from other athletic and nonathletic populations based on physiology, demographics (eg, all biologically male players), and the number of years participating in sport, among other factors. As a result, these results should not be extrapolated to populations other than former NFL players; however, future authors should consider replicating this approach in various athletic populations.

Former NFL players with a history of lower extremity injury, surgery, or OA had a higher prevalence of CRFs than former players with no such history. The association between lower extremity injury, surgery, or OA and the prevalence of CRFs was stronger among individuals <50 years of age. These findings provide insight regarding the pathway to chronic diseases that may be initiated by joint injury early in life, even among individuals who participate in elite sport.

Supplemental Table 1. Adjusted Prevalence Ratios (PRs) for Individual Cardiovascular Disease Risk Factors (CRFs).

Supplemental Table 2. Characteristics of Participants With Knee-Related Injury or Surgery.

Supplemental Table 3. Prevalence Ratios (PRs) for Predictors of Cardiovascular Disease in Participants With Knee-Related Injury or Surgery History or No Reported Injury or Surgery History (n = 1486).

Supplemental Table 4. Prevalence Ratios (PRs) for Predictors of Reporting One or More Cardiovascular Disease Risk Factors in Participants With Knee-Related Injury or Surgery History or No Reported Injury or Surgery History (n = 1486).

Found at DOI: http://dx.doi.org/10.4085/1062-6050-0437.22.S1